BPM Proliferation Drives the Need for Process Intelligence

Many BPMS solutions are falling short of their promise to give organizations an end-to-end view of business processes. In most organizations, processes live in a variety of business systems within the enterprise. Thus, organizations will inevitably have to contend with the challenge of managing multiple process silos. Consequently, "process intelligence" will quickly prove to be an even more strategic technology than BPM suites for the "process-aware" organizations striving to improve their business performance.

Organizations today are more "process aware" then ever before. Many, if not most, companies have active projects underway to improve their business processes. Often, a variety of technology is deployed to further the advancement of business process management and improvement. In recognizing the need to better manage processes, there are a number of common requirements that organizations articulate. Frequently, what is heard includes:

• We need a better way to understand why business objectives are not being met.
• We need an end-to-end view of business processes and an ability to align cross-functional activities.
• We need to understand the impact of change on a given process before the change takes place.

While traditional approaches to business process management (BPM) technology aspire to address these needs, they fall short in several areas. BPM suite (BPMS) technology often does a good job of providing management and optimization for the processes that it directly manages. However, most BPM suites are not designed to address the management of processes not directly within their control.

At their core, BPM suites provide an infrastructure for applications. While they are an incredibly innovative way of delivering agile applications, they are application infrastructures nonetheless. This is an inherently limiting approach.

In most organizations, processes live in a variety of business systems within the enterprise, including enterprise resource planning (ERP), enterprise content management (ECM) and specialized line-of-business systems supporting everything from mortgage and policy processing to IT service requests. It is impractical to replace these systems with BPMS technology, and neither cost effective nor technologically effective to "wrap" each of these systems with a BPMS when there is BPM-like technology already inside many of them.

In fact, BPM technology is becoming all the more pervasive, being built into application servers, portals, application suites and development tool kits. BPM silos are developing within organizations due to this technology proliferation. As organizations progress along the BPM maturity curve (deploying more and more BPM technology) they realize what they really need is information about their processes, not an increase in their process application infrastructure.

At this point, many organizations may consider leveraging business intelligence (BI) software to solve the issue of providing better process information. However, a quick evaluation finds traditional BI better suited for reporting on data (purchase orders, invoices) rather than processes (process cycle time, wait times, working times, resources and process costs). This is because event data from a BPMS carries with it a great deal of process-specific context.

For example, to accurately calculate the number of days a particular work item took to complete, the analytic engine must understand the concept of a "business day" and be able to adjust for holidays, vacations, time zones, etc. The fact that many organizational processes are enabled by more than one BPMS or application presents the even larger challenge of event data normalization. Moreover, BI systems were designed—and are still limited today—to providing largely historical information, with minimal real time or proactive capabilities. Finally, BI is notoriously project-intensive, requiring significant effort to properly design the underlying data structure (often termed "data warehouse") as well as the associated reports and dashboards typically required by most organizations. BI projects for BPMS analysis are even more difficult when the goal is to structure a data warehouse for process event-specific data that must be derived from a variety of incongruent BPMS and application technologies.

BPM and BI are, in fact, two very different technologies and were developed to stand on their own, delivering value in disparate ways. However, more and more organizations are realizing the value of bringing these two technologies together. Pairing up BI and BPM helps shore up the technologies' respective shortcomings. As a result, the aggregation of data that results from the combination of these technologies can best be described as "process intelligence."

Process Intelligence

Process intelligence suggests a broader reach across systems and processes that may be ancillary to the BPMS, or powered by other suites or enterprise applications.

Process intelligence is thus global, providing end-to-end visibility and control of cross-functional business processes, which may be supported by a number of disparate technology infrastructures.

Process intelligence excels at complex, human-centric BPM involving thousands of participants and millions of transactions on a daily basis. Common scenarios involve customer enrollment, loan origination and claims processing. In some cases, the complexity of the process involves coordination of work among several different functional areas including customer service, underwriting and new business processing.

Once these processes are successfully automated, customers begin their continuous pursuit of year-over-year ROI. Initially, that ROI is largely derived from the implementation of the software itself and the savings associated with automation. However, in order to realize continuing ROI through decreasing operational costs and call wait times, increased customer satisfaction and profit margins, etc., customers must maintain their vigilance for optimization. All this requires extensive